1 Overview

The Australian National University (ANU) is a large, complex organization that faces many risks. For example, its 2020 Annual Report describes the extensive financial resources that it enjoys as well as the uncertain obligations, or risks, that it faces. In this report, we document loss experience of risks that ANU has historically found to be insurable.

1.1 What is an Insurable Risk?

A risk is said to be insurable if it potentially can be transferred to another party for a fee. Commercially available insurance is an important mechanism for doing so but firms also employ other options such as self insurance pools, peer to peer risk exchanges, and so on.

One way to get insights into whether a risk is insurable is to consider the many risks faced by an organization; these can be described based on concerns expressed by risk managers. To this end, we cite a survey of global risk managers conducted by Aon in 2019, Solutions (2019) (for another viewpoint, see a Deloitte survey, Insights (2019)). From the survey (Page 13), “… Each year we offer respondents the chance to assess their future risk landscape and project the top five risks that their organizations will face in three years’ time.” The survey identifies 69 risks, some of which are insurable, some partially insurable, and others uninsurable. For example, the top five risks are:

  1. Economic slowdown/ slow recovery - uninsurable
  2. Damage to reputation/brand - partially insurable
  3. Accelerated rates of change in market factors - uninsurable
  4. Business interruption - insurable
  5. Increasing competition - uninsurable

This survey was conducted in mid-2019 before the onslaught of the COVID pandemic. Interestingly, global risk managers only identified “Pandemic risk/ health crises” as rank number 60 risk factor. In this report, we focus on risks that a manager might transfer to another entity and so highlight insurable risks. From the survey, the top insurable risks (with their rank in the survey) are given in Table 1.1.

Table 1.1. Top Insurable Risks Facing Firms (Source: Solutions (2019))

\[ \small{ \begin{matrix} \begin{array}{l|l} \hline \hline \textbf{Risk} & \textbf{Risk} \\ \hline \text{4. Business interruption} & \text{44. Directors and Officers personal liability}\\ \text{19. Counter-party credit risk} & \text{47. Fraud}\\ \text{21. Property damage} & \text{52. Theft}\\ \text{22. Environmental risk} & \text{55. Terrorism sabotage}\\ \text{23. Weather natural disasters} & \text{56. Safety and Pharmacovigilance}\\ \text{24. Third party liability} & \text{61. Harassment discrimination}\\ \text{28. Injury to workers} & \text{66. Kidnap and ransom}\\ \text{40. Product recall} & \text{67. Extortion}\\ \hline \hline \end{array} \end{matrix} } \]

1.2 ANU Approach to Managing Insurable Risks

Like many large organizations, ANU has extensive risk control activities in place focusing on risk avoidance, loss prevention or reduction, and so-called “risk control transfer” (whereby the responsibility of the risk may be transferred). These activities are designed to mitigate the financial impact that a risk can have on the organization. Anticipating and reducing the potential impact of risks are critical activities of risk managers and are not described further in this report. You can learn more about ANU’s risk philosophy in an overview of ANU Risk Management Policy Statement, ANU (2022a).

Nonetheless, even with the best risk management processes, losses do occur and risk financing methods are needed to provide resources for reimbursing the cost of a loss. These methods fall into two broad classes: retention and risk transfer. ANU takes a layered approach to financing risks:

  • At the lowest level, individual business units are responsible for handling small adverse outcomes.
  • For small and intermediate outcomes, ANU has created a Self-Insurance Reserve (SIR) Pool.
  • For larger adverse outcomes, ANU transfers risks to external insurance companies (for a price).

The amount of risk retention and transfer depends upon the risk type and will be covered throughout this report.

1.3 Report Structure

The report is organized as follows:

  • The SIR pool is owned and operated by ANU. This entity is further described in Section 2 together with a summary of its loss experience.
  • The relationships among external insurers, and loss experience, is more complex. This is the subject of Section 3.
  • A type of risk sitting outside of this layered structure is workers’ compensation. This has a separate history and claims experience that is described in Section 4.
  • Section 5 provides access to the data underpinning this report.

A supplemental appendix Section 6 provides a brief overview of the many insurance coverages considered in this report.

2 Self-Insurance Reserve Pool

Major universities such as ANU organize budget responsibilities by layers, beginning from smaller units such as colleges and departments. In addition, ANU owns or is affiliated with several other organizations such as ANU Enterprise Pty Ltd, ANU Union, and so forth. (ANU (2018) provides a list of “named insureds” organizations that participate.) Rather than have each financially responsible unit purchase insurance according to their own needs, in 1994 ANU organized the so-called “Self-Insurance Reserve” or SIR for short. From ANU (2022b), the Self Insurance Reserve (SIR) pool is “for insurance losses that except for the policy excess or coverage limit, would otherwise be covered under the commercial insurance program.” In this way, the university facilitates economies of scale and reduces costs of insurance purchases.

On the one hand, this mechanism affects incentives of units within ANU and so is important. See the 2018-2019 Self-Insurance Reserve Policy (ANU (2018)) for more information about the structure and operations of the SIR pool. On the other hand, because the SIR pool is owned and operated by ANU, there is no external transfer of risk. For many risk financing purposes, this level of detail can be ignored.

2.1 Loss Summary

To describe the magnitude of the SIR pool, we start with Table 2.1 that provides the number of losses and loss amounts over years 2012-2020. As is common when the numbers of loss are small, the experience is quite variable from year to year.

InsSIRSummaryClaims <- read.csv("..\\ANUData\\SIRClaims.csv", header = T)

TableSIR1 <- summaryBy(Amount ~ Year, data = InsSIRSummaryClaims, FUN = function(x) {
    c(num = length(x), m = mean(x), s = sum(x))
})
T1 <- colSums(TableSIR1)
TableSIR <- rbind(TableSIR1, T1)
TableSIR[9, 3] <- TableSIR[9, 4]/TableSIR[9, 2]
names(TableSIR) <- c("Year", "Number", "Average Severity", "Total Loss")
TableSIR <- round(TableSIR, digits = 0)
TableSIR[9, 1] <- "Total"
kableExtra::kbl(TableSIR, caption = "**Summary of SIR Claims, 2012-2020**", align = "ccrr",
    format.args = list(big.mark = ",", scientific = FALSE), table.attr = "style='width:80%;'") %>%
    kableExtra::kable_classic(full_width = T, html_font = "Cambria") %>%
    kable_styling(bootstrap_options = c("striped", "condensed"))
Table 2.1: Summary of SIR Claims, 2012-2020
Year Number Average Severity Total Loss
2012 5 75,002 375,012
2013 2 9,541 19,082
2014 11 17,453 191,981
2015 4 68,826 275,305
2016 12 45,952 551,422
2017 7 7,465 52,253
2018 5 32,324 161,619
2019 6 1,184 7,105
Total 52 31,419 1,633,779

To get a sense of the distribution, Figure 2.1 provides boxplots of individual losses by year. The left-hand panel shows the annual distributions in the original units (AUD) and the right-hand panel gives each annual distribution but on a logarithmic scale. Both plots exhibit substantial variation over time.

p1 <- ggplot(data = InsSIRSummaryClaims, aes(x = factor(Year), y = Amount)) + geom_boxplot() +
    theme_bw() + xlab("Year") + theme(axis.text.x = element_text(size = 8)) + ylab("Loss Amount")
p2 <- ggplot(data = InsSIRSummaryClaims, aes(x = factor(Year), y = Amount)) + geom_boxplot() +
    theme_bw() + xlab("Year") + scale_y_continuous(trans = "log10") + theme(axis.text.x = element_text(size = 8)) +
    ylab("Loss Amount")
grid.arrange(p1, p2, nrow = 1)
**Distribution of Self-Insurance Reserve Pool Losses by Year**

Figure 2.1: Distribution of Self-Insurance Reserve Pool Losses by Year

2.2 SIR Deductibles

SIR losses over time are subject to deductibles. To interpret this historical information, one needs to understand deductibles that have been applied.

From the 2018-19 Self Insurance Reserve Policy (ANU (2018)), we have information about deductibles for that policy year, including external insurance (see Section 3 for further discussion on external insurance). Deductibles for policy year 2020-2021 are largely consistent with prior years, except that (internal) SIR deductibles for library (part of property), crime, cyber, statutory liability, motor vehicle, group personal accident, corporate travel, marine hull, and marine cargo have been dropped. Beginning in policy year 2021-2022, the SIR property deductible for damage to a building structure has also been waived.

Self-Insurance Reserve Excess Limits 2018/2019

2.3 Detailed SIR Pool Data

Deductibles and upper limits vary by the type of risk so each loss is categorized accordingly.

InsSIRSummaryClaims <- read.csv("..\\ANUData\\SIRClaims.csv", header = T)
by_Year_Cat <- InsSIRSummaryClaims %>%
    group_by(Year, Category)
CrossTabs1 <- by_Year_Cat %>%
    summarise(n = n()) %>%
    spread(Year, n)
CrossTabs1[is.na(CrossTabs1)] = 0

CrossTabs1A <- data.frame(CrossTabs1)
CrossTabs1A1 <- CrossTabs1A[, -1]
T21col <- colSums(CrossTabs1A1)
T21row <- rowSums(CrossTabs1A1)
Total <- c(T21row, sum(T21col))

CrossTabs1B <- rbind(CrossTabs1, c(NA, T21col))
CrossTabs1B1 <- cbind(CrossTabs1B, Total)
if (CAMPUS == FALSE) {
    levels(CrossTabs1B1$Category)[is.na(CrossTabs1B1$Category)] <- "Total"
}
CrossTabs1B1$Category[is.na(CrossTabs1B1$Category)] <- "Total"

kableExtra::kbl(CrossTabs1B1, caption = "Count of SIR Claims by Category and Year", align = "lcccccccccccc",
    table.attr = "style='width:80%;'") %>%
    kableExtra::kable_classic(full_width = F, html_font = "Cambria") %>%
    kable_styling(bootstrap_options = c("striped", "condensed"))
Table 2.2: Count of SIR Claims by Category and Year
Category 2012 2013 2014 2015 2016 2017 2018 2019 Total
Corporate Travel 0 0 1 0 0 0 1 0 2
Crime 0 0 2 2 1 1 0 0 6
Group Personal Accident 0 0 1 0 0 0 0 2 3
Motor Vehicle 0 0 3 0 5 1 1 4 14
Property: Buildings / Contents 3 2 3 2 6 5 3 0 24
Public Liability 2 0 1 0 0 0 0 0 3
Total 5 2 11 4 12 7 5 6 52 
CrossTabs2 <- by_Year_Cat %>%
    summarise(Loss_Amount = sum(Amount)) %>%
    spread(Year, Loss_Amount)
CrossTabs2[is.na(CrossTabs2)] = 0
CrossTabs2A <- data.frame(CrossTabs2)
CrossTabs2A1 <- CrossTabs2A[, -1]
T22col <- colSums(CrossTabs2A1)
T22row <- rowSums(CrossTabs2A1)
Total <- c(T22row, sum(T22col))

CrossTabs2B <- rbind(CrossTabs2, c(NA, T22col))
CrossTabs2B2 <- cbind(CrossTabs2B, Total)
if (CAMPUS == FALSE) {
    levels(CrossTabs2B2$Category)[is.na(CrossTabs2B2$Category)] <- "Total"
}
CrossTabs2B2$Category[is.na(CrossTabs2B2$Category)] <- "Total"

kableExtra::kbl(CrossTabs2B2, caption = "**Sum of SIR Claims by Category and Year**",
    align = "lrrrrrrrrrr", format.args = list(big.mark = ",", scientific = FALSE),
    table.attr = "style='width:80%;'") %>%
    kableExtra::kable_classic(full_width = F, html_font = "Cambria") %>%
    kable_styling(bootstrap_options = c("striped", "condensed"))
Table 2.3: Sum of SIR Claims by Category and Year
Category 2012 2013 2014 2015 2016 2017 2018 2019 Total
Corporate Travel 0 0 300 0 0 0 650 0 950
Crime 0 0 20,465 48,583 73,748 3,416 0 0 146,212
Group Personal Accident 0 0 10,000 0 0 0 0 1,692 11,692
Motor Vehicle 0 0 5,750 0 16,512 760 1,605 5,413 30,040
Property: Buildings / Contents 291,679 19,082 152,966 226,722 461,162 48,077 159,364 0 1,359,052
Public Liability 83,333 0 2,500 0 0 0 0 0 85,833
Total 375,012 19,082 191,981 275,305 551,422 52,253 161,619 7,105 1,633,779

You can develop a feel for how these losses arise through a cursory examination of individual SIR claims.

Individual SIR Claims

Data access is available in Section 5.1.

3 Insurance Data

3.1 Overview of Insurance Programs

To get a sense of the broad risk financing program that ANU employs, Table 3.1 summarizes 15 coverages and premiums paid in year 2020-2021. This summary is based on information provided by Gallagher, a brokerage firm retained by ANU. Due to confidentiality of policies, Table 3.1 excludes the following coverages:

  • Directors & Officers
  • Excess Directors & Officers
  • Directors & Officers Supplementary Legal Expenses.

The insurance program is important to the overall financial health of ANU. According to the 2020 Annual Report (page 96), total expenditures for ANU in 2020 were $1,315 million. Thus, the $24,407,255 in insurance premiums represents about 1.86% of total expenditures.

As is evident from Table 3.1, the property risk is by far the most important, accounting for about 95% of the premium. Notably, the property deductible is currently $5 million, representing a large uninsured risk. The second most important risk type is General and Products Liability, representing $448,500 in premiums or about 2% of the total. The other 13 coverages sum to $692,056 that represents about 3% of the total.

Table 3.1: ANU Insurance Premiums, 2020-2021
Class.of.Insurance Insurer Limit Deductible Premium
Property London Syndicate and Others 1,000,000,000 5,000,000 23,564,759
General and (G & P) Products Liability Newline 20,000,000 100,000 340,000
G & P Umbrella Liability Liberty 50,000,000 20,000,000 27,500
G & P 1st Exess Liability QBE 100,000,000 50,000,000 27,500
G & P 2nd Excess Liability Chubb 150,000,000 100,000,000 17,500
G & P 3rd Excess Liability CGU 200,000,000 150,000,000 16,000
G & P 4th Excess Liability Zurich 250,000,000 200,000,000 20,000
Cyber London 2,000,000 250,000 75,721
Crime AIG 20,000,000 100,000 100,000
Employment Practices Liability AIG 2,000,000 100,000 84,000
Expat - December Renewal Chubb 11,676
Group Personal Accident Chubb As per schedule Various 104,920
Marine Cargo Richard Oliver (QBE) 5,000,000 5,000 6,127
Marine Hull Richard Oliver (QBE) 5,000,000 150 11,552
Motor Vehicle Vero As per schedule 1,000 84,700
Professional Indemnity Newline $20m / $40m 100,000 130,000
Medical Malpractice Newline $20m / $40m 100,000
Clinical Trial Newline $20m 2,500
Statutory Liability Berkley Insurance Australia (SUA) 1,000,000 $1,000 / $15,000 8,360
Travel Chubb As per schedule Various 75,000
TOTALS 24,407,255

You can learn more about the coverages in Appendix Section 6.This coverage data applies to the 2020-21 policy year. (Note: For the 2021-2022 policy, ANU decided to self-fund the cyber risk due to the lack of availability of coverage at an affordable price.)

Prior to 31 October 2020, several of ANU’s risk exposure were covered by Unimutual. Unimutual, formed in 1989, is an insurance mutual pool whose members consist of Australian universities, higher education providers, and associated entities. As of the end of 2020, there were 26 universities and 27 non-University (“Allied”) members. Like many insurance pools, it was formed because commercial insurers did not provide coverage for risks faced by a specific marketplace, in this case the higher education sector.

Unimutual classifies risks according to five types:

  • Liability Protection (including General and Products Liability, Professional Liability, Malpractice, and Clinical Trials Protection)
  • Property Protection
  • Cyber Protection
  • Environmental Liability
  • Active Assailant.

The Unimutual website provides additional details.

3.2 Recent Claim Data

We start with a summary of recent insurance claims data in Table 3.2. By tradition, the insurance policy years begin on 1 November and finish 31 October. So, for example, Num17.18 refers to the number of claims between 1 November 2017 and 31 October 2018, inclusive. (Note that Expatriate claims are not included as they are not part of the Gallagher report).

In subsequent sections, we analyze detailed claims data where available.

Table 3.2: ANU Insurance Claims Statistics, 2017-2020
Coverage Num19.20 Num18.19 Num17.18 Amt19.20 Amt18.19 Amt17.18
Travel 255 201 139 363,441 192,405 261,155
Personal Accident 2 17 16 3,601 22,259 16,747
Motor Vehicle 118 26 28 1,247,827 57,880 49,697
General & Product Liability 2 1 1 13,993 3,309 520,000
Property 1 0 1 249,500,000 0 49,500,000
Cyber 0 1 0 0 1,650,364 0
Employment Practices Liability 5 1 3 73,475 52,576 82,680
Marine Hull 0 0 1 0 0 14,084


Most notable in Table 3.2 are the following:

  • Property claim in 2019-2020 for $249.5 million, nearly a quarter of a billion dollars. This was caused by a severe hailstorm on 20 January 2020 (see the Motor Vehicle claims in the same year). To put this in perspective, total ANU assets in 2020 was $4.66 billion and equity (net assets which is total assets minus liabilities) was $2.65 billion. Thus, the single claim represents 5.35% of total assets and 9.4% of equity, both significant fractions.
  • Property claim in 2017-18 for $49.5 million. This was caused by a flooding incident on 25 February 2018.
  • Cyber claim of $1,650,364. This was caused by a data breach in late-2018.

3.3 Detailed Data by Lines of Insurance

3.3.1 Corporate Travel

Universities purchase corporate travel policies to cover employees and students traveling on official university business for a wide variety of accidents and incidents while away from the campus or primary workplace. This broad coverage includes medical care and evacuation, loss of personal property, extraction for political and weather related reasons, and more. Additional details can be found in ANU’s corporate travel policy. You can also learn more about this line of business from ANU’s insurer, Chubb Travel:

The Data

The data provided are maintained by the insurer, Chubb. These data were accessed on 29 July 2022. Compared to other coverages, the data history is long and stable. This coverage began on 1 November 2006. See the following count of claims.

TravelClaims <- read.csv("..\\ANUData\\TravelClaims2022.csv", header = T)
tableTravel <- t(table(TravelClaims$UW.Year))

kableExtra::kbl(tableTravel, caption = "**Travel Claims Frequency** ") %>%
    kableExtra::kable_classic_2(position = "center")
Table 3.3: Travel Claims Frequency
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
41 74 102 166 158 141 143 161 158 158 154 139 205 274 1 32

From this data set, there are 2107 incurred claims. Of these claims, there are 269 zeros and an additional 3 claims where the incurred claim is less than 10. We omit these claims in our analysis.

Details of the Travel Claims Data

In addition to the data provided in Section 5.2, ANU analysts have access to information such as descriptions of losses. To protect the privacy of claimants, many of whom are students, we do not provide details of this information. However, to give you a feel for the kind of information available, the following gives a summary.

Description of Loss Summary Information

Analysis of Incurred Losses

There are 1835 incurred losses. The distribution of incurred losses is stable over time.

TravelClaims1 <- subset(TravelClaims, Incurred.Loss > 10)

ggplot(data = TravelClaims1, aes(x = factor(UW.Year), y = Incurred.Loss)) + geom_boxplot() +
    theme_bw() + xlab("Year") + scale_y_continuous(trans = "log10")
**Distribution of Travel Losses by Year**

Figure 3.1: Distribution of Travel Losses by Year 

sumTravelClaims <- t(summary(TravelClaims1$Incurred.Loss))
kableExtra::kbl(sumTravelClaims, caption = "**Travel Claims Summary Statistics** ") %>%
    kableExtra::kable_classic_2(full_width = F, position = "center")
Table 3.4: Travel Claims Summary Statistics
Min. 1st Qu. Median Mean 3rd Qu. Max.
11.32 255.26 550 1731.43251 1298.67 422603.4

One can fit a distribution to the losses. In the following, we fit via maximum likelihood the gamma, Pareto, and lognormal distributions to incurred losses. Figure 3.2 suggests that the lognormal distribution appears to be the best fit. If you would like background on this type of analysis, see Actuarial Community (2020).

library(VGAM)
x <- seq(0, 15, by = 0.01)

# Inference assuming a gamma distribution
fit.gamma <- vglm(Incurred.Loss ~ 1, family = gamma2, data = TravelClaims1)
theta <- exp(coef(fit.gamma)[1])/exp(coef(fit.gamma)[2])  # theta = mu / alpha
alpha <- exp(coef(fit.gamma)[2])
fgamma_ex <- dgamma(exp(x), shape = alpha, scale = theta) * exp(x)

# Pareto
fit.pareto <- vglm(Incurred.Loss ~ 1, paretoII, loc = 0, data = TravelClaims1)
fpareto_ex <- dparetoII(exp(x), loc = 0, shape = exp(coef(fit.pareto)[2]), scale = exp(coef(fit.pareto)[1])) *
    exp(x)

# Lognormal
fit.LN <- vglm(Incurred.Loss ~ 1, family = lognormal, data = TravelClaims1)
flnorm_ex <- dlnorm(exp(x), mean = coef(fit.LN)[1], sd = exp(coef(fit.LN)[2])) *
    exp(x)

plot(density(log(TravelClaims1$Incurred.Loss)), main = "", xlab = "Log Claims")  #
# , ylim = c(0 ,0.37) )

lines(x, fgamma_ex, col = "blue", lty = 2)
lines(x, fpareto_ex, col = "purple", lty = 3)
lines(x, flnorm_ex, col = "green", lty = 3)
legend("topleft", c("log(Claim)", "Gamma", "Pareto", "Lognormal"), cex = 0.8, lty = c(1,
    2, 3, 4, 1, 1), col = c("black", "blue", "purple", "green"))  #4 in lty is 'longdash'
**Distribution of Travel Losses with Superimposed Fitted Distributions**

Figure 3.2: Distribution of Travel Losses with Superimposed Fitted Distributions


There is a slight increase in claims over time, about 3.8% per year. As a follow-up, readers may wish to think about mechanisms for adjusting claims for inflation.

Regression Fit of Travel Claims Trend

Claim Frequency

Clm_Count <- as.numeric(table(TravelClaims$UW.Year))
N = sum(Clm_Count)
lambdahat = sum(Clm_Count)/length(Clm_Count)
loglikNB <- function(parms) {
    r = parms[1]
    beta = parms[2]
    llk <- -sum(log(dnbinom(Clm_Count, size = r, mu = r * beta)))
    llk
}

ini.NB <- c(10, 5)
# loglikNB(ini.NB )
zop.NB <- nlminb(ini.NB, loglikNB, lower = c(0.000001, 0.000001), upper = c(Inf,
    Inf))

rhat.NB = zop.NB$par[1]
betahat.NB = zop.NB$par[2]

The number of claims are sufficient that a separate frequency model could be considered. For the frequency of claims, there are 2107 claims over the 2006-2021 period that amounts to 131.69 per year. One might assume that annual claims can be fit using a single distribution to the entire period, such as a Poisson or a negative binomial. Another option is to fit a distribution starting in years 2009, where this is an increase in the amount of claims from prior years. A third option is to omit experience from underwriting year 2019 and on where the number of claims fluctuated dramatically, in part due to the Covid epidemic.

To start, we fit a Poisson distribution and a negative binomial distribution to all claims. The resulting maximum likelihood estimators are \(\hat{\lambda}=\) 131.69 for the Poisson and \(\hat{r}=\) 1.75 and \(\hat{\beta}=\) 75.16 for the negative binomial. The following table compares the empirical percentiles to those under the Poisson and negative binomial. Both fitted distributions did well and neither outperformed the other.

Claim Frequency Fit

The mechanism for reporting no claims is uncertain, so probably the best option is to remove the zeroes entirely and fit zero-truncated distributions, see for example, Section 2.5.1 of Loss Data Analytics. We leave this as an exercise for motivated readers.

Data access is available in Section 5.2.

3.3.2 Group Personal Accident

Group personal accident insurance offers financial protection in case of injury or death resulting from an incident that occurs on the job. Like workers’ compensation to be described in Section 4, group personal accident offers insurance coverage and liability insurance protection against accidental death or injury. Unlike workers’ compensation, group personal accident covers students and ANU’s voluntary workers.

Several limits apply including $1,000,000 for the period of insurance, $600,000 for non-scheduled flights, and others. These limits were not reached in the data we consider.

The Data

The data provided to us are maintained by the insurer, Chubb. These data were accessed on 29 July 2022. These data began in underwriting year 2007. See the following count of claims.

GPAClaims <- read.csv("..\\ANUData\\GroupPersonalAccidentClaims2022.csv", header = T)
tablePersonAcc <- t(table(GPAClaims$UW.Year))
GPAClaimsGT0 <- subset(GPAClaims, Incurred.Loss > 0)

kableExtra::kbl(tablePersonAcc, caption = "**Group Personal Accident Claims Frequency** ") %>%
    kableExtra::kable_classic_2(position = "center")
Table 3.5: Group Personal Accident Claims Frequency
2007 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
2 3 5 6 12 14 5 11 17 17 10 35 11

From this data set, there are 148 incurred claims. Of these claims, there are 35 zeros and an additional 0 claims where the incurred claim is less than 10. We omit these claims in our analysis.

Analysis of Incurred Losses

For this coverage, there is a “7 day excess” for weekly benefits but none for general benefits. The database documentation provided to us, and the data we provide, do not indicate whether the excess has been triggered; we have only paid claims. Because of the relatively small size of this class of insurance, we ignore the effects of deductibles for this line.

There are 112 incurred losses. Figure 3.3 indicates that the incurred losses are stable over time.

ggplot(data = GPAClaimsGT0, aes(x = factor(UW.Year), y = Incurred.Loss)) + geom_boxplot() +
    theme_bw() + xlab("Year") + scale_y_continuous(trans = "log10")
**Distribution of Group Personal Accident Losses by Year**

Figure 3.3: Distribution of Group Personal Accident Losses by Year 

sumGPAClaimsGT0 <- t(summary(GPAClaimsGT0$Incurred.Loss, digits = 0))
kableExtra::kbl(sumGPAClaimsGT0, caption = "**Personal Accident Incurred Losses** ") %>%
    kableExtra::kable_classic_2(full_width = F, position = "center")
Table 3.6: Personal Accident Incurred Losses
Min. 1st Qu. Median Mean 3rd Qu. Max.
90 500 1000 2000 2000 30000

One can fit a distribution to the losses. In the following, we fit via maximum likelihood the gamma, Pareto, and lognormal distributions to incurred losses. Figure 3.4 suggests that the lognormal distribution appears to be the best fit.

library(VGAM)
x <- seq(0, 15, by = 0.01)

# Inference assuming a gamma distribution
fit.gamma <- vglm(Incurred.Loss ~ 1, family = gamma2, data = GPAClaimsGT0)
theta <- exp(coef(fit.gamma)[1])/exp(coef(fit.gamma)[2])  # theta = mu / alpha
alpha <- exp(coef(fit.gamma)[2])
fgamma_ex <- dgamma(exp(x), shape = alpha, scale = theta) * exp(x)

# Pareto
fit.pareto <- vglm(Incurred.Loss ~ 1, paretoII, loc = 0, data = GPAClaimsGT0)
fpareto_ex <- dparetoII(exp(x), loc = 0, shape = exp(coef(fit.pareto)[2]), scale = exp(coef(fit.pareto)[1])) *
    exp(x)

# Lognormal
fit.LN <- vglm(Incurred.Loss ~ 1, family = lognormal, data = GPAClaimsGT0)
flnorm_ex <- dlnorm(exp(x), mean = coef(fit.LN)[1], sd = exp(coef(fit.LN)[2])) *
    exp(x)

plot(density(log(GPAClaimsGT0$Incurred.Loss)), main = "", xlab = "Log Claims", ylim = c(0,
    0.4))

lines(x, fgamma_ex, col = "blue", lty = 2)
lines(x, fpareto_ex, col = "purple", lty = 3)
lines(x, flnorm_ex, col = "green", lty = 3)
legend("topleft", c("log(Claim)", "Gamma", "Pareto", "Lognormal"), cex = 0.8, lty = c(1,
    2, 3, 4, 1, 1), col = c("black", "blue", "purple", "green"))
**Distribution of Group Personal Accident Losses with Superimposed Fitted Distributions**

Figure 3.4: Distribution of Group Personal Accident Losses with Superimposed Fitted Distributions


The following analysis indicates that there is no appreciable trend in claims severity over time.

Regression Fit of Personal Accident Trend

Data access is available in Section 5.3.

3.3.3 Motor Vehicle

This policy covers ANU’s vehicles including cars, vans, utilities, and motorcycles.

There are two parts to this coverage, one for comprehensive damage to the insured vehicles and a second for legal liability.

Part 1 - Loss or Damage to Insured Vehicles. This includes:

  • Market Value for Cars, vans, utilities, motorcycles, or 4WD’s of less than two tonne carrying capacity.
  • All other vehicle types: Market Value or the amount shown in the agreed schedule of vehicles, whichever is less.

Part 2 - Legal Liability. This includes:

  • $50 Million for all claims arising from the one accident or series of accidents resulting from the one original cause.
  • $1 Million for all claims arising from the one accident or series of accidents resulting from the one original cause where the vehicle is used for transportation of dangerous goods.

The excesses are cumulative and apply to all claims.

For each event, or series of events arising from the one originating cause, ANU bears the amount of the excess in respect of each and every insured vehicle, unless stated otherwise.

  • Basic Excess: $1,000
  • Mobility Scooter Excess: $500
  • Event Excess: $50,000

The Data

The data provided to us are maintained by the insurer, Vero Insurance Limited. These data were accessed on 8 August 2022. These data began in underwriting year 2012. See the following count of claims.

AutoClaims <- read.csv("..\\ANUData\\MotorClaims2022.csv", header = T)
UwYear <- as.Date(AutoClaims$Policy.Term.Start.Date, "%d/%m/%Y")
AutoClaims$UW.Year <- as.numeric(format(UwYear, format = "%Y"))
tableAutoClaims <- t(table(AutoClaims$UW.Year))
kableExtra::kbl(tableAutoClaims, caption = "**Motor Vehicle Claim Frequency** ") %>%
    kableExtra::kable_classic_2(position = "center")
Table 3.7: Motor Vehicle Claim Frequency
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
11 20 32 15 19 30 28 26 120 6 11

From this data set, there are 318 incurred claims. Of these claims, there are 50 zeros and an additional 0 claims where the incurred claim is less than 10. We omit these claims in our analysis.

Details of the Motor Claims Data

ANU analysts have access to additional information such as descriptions of losses. To protect the privacy of claimants, many of whom are students, we do not provide details of this information. However, to give you a feel for the kind of information available, the following gives a summary.

Description of Loss Type Information

Analysis of Incurred Losses

The data provided to us contain the excess as well as the amount paid by Vero. In the following, we provide a basic analysis ignoring the effects of deductibles/excess. We recommend that motivated readers extend our analysis to account for this deductible in both the severity and frequency.

There are 268 incurred losses. When examining data over time, we see that experience in 2020 was dramatically different both in the number of claims and the severity of claims. As discussed, this was due to a large hail storm. For the moment, we treat these claims as part of a single distribution, although other approaches could be considered.

ggplot(data = AutoClaims1, aes(x = factor(UW.Year), y = Motor.Net.Incurred)) + geom_boxplot() +
    theme_bw() + xlab("Year") + scale_y_continuous(trans = "log10")
**Distribution of Motor Vehicle Losses by Year**

Figure 3.5: Distribution of Motor Vehicle Losses by Year 

sumAuto <- t(summary(AutoClaims1$Motor.Net.Incurred, digits = 0))
kableExtra::kbl(sumAuto, caption = "**Motor Vehicle Incurred Losses** ") %>%
    kableExtra::kable_classic_2(full_width = F, position = "center")
Table 3.8: Motor Vehicle Incurred Losses
Min. 1st Qu. Median Mean 3rd Qu. Max.
20 1000 3000 7000 9000 60000

One can fit a distribution to the losses. In the following, we fit via maximum likelihood the gamma, Pareto, and lognormal distributions to incurred losses. The lognormal distribution appears to be the best fit.

library(VGAM)
x <- seq(0, 15, by = 0.01)

# Inference assuming a gamma distribution
fit.gamma <- vglm(Motor.Net.Incurred ~ 1, family = gamma2, data = AutoClaims1)
theta <- exp(coef(fit.gamma)[1])/exp(coef(fit.gamma)[2])  # theta = mu / alpha
alpha <- exp(coef(fit.gamma)[2])
fgamma_ex <- dgamma(exp(x), shape = alpha, scale = theta) * exp(x)

# Pareto
fit.pareto <- vglm(Motor.Net.Incurred ~ 1, paretoII, loc = 0, data = AutoClaims1)
fpareto_ex <- dparetoII(exp(x), loc = 0, shape = exp(coef(fit.pareto)[2]), scale = exp(coef(fit.pareto)[1])) *
    exp(x)

# Lognormal
fit.LN <- vglm(Motor.Net.Incurred ~ 1, family = lognormal, data = AutoClaims1)
flnorm_ex <- dlnorm(exp(x), mean = coef(fit.LN)[1], sd = exp(coef(fit.LN)[2])) *
    exp(x)

plot(density(log(AutoClaims1$Motor.Net.Incurred)), main = "", xlab = "Log Claims",
    ylim = c(0, 0.37))

lines(x, fgamma_ex, col = "blue", lty = 2)
lines(x, fpareto_ex, col = "purple", lty = 3)
lines(x, flnorm_ex, col = "green", lty = 3)
legend("topleft", c("log(Claim)", "Gamma", "Pareto", "Lognormal"), cex = 0.8, lty = c(1,
    2, 3, 4, 1, 1), col = c("black", "blue", "purple", "green"))
**Distribution of Motor Vehicle Losses with Superimposed Fitted Distributions**

Figure 3.6: Distribution of Motor Vehicle Losses with Superimposed Fitted Distributions


There is a strong trend in motor vehicle claims severity over time. Will need to assess the impact of the recent hail event in 2020.

Regression Fit of Motor Vehicle Trend

Data access is available in Section 5.4.

3.3.4 Employment Practices Liability

Employment practices liability is a type of liability for wrongful acts arising from the employment process. The insurance covers claims by workers that their legal rights as employees of the company have been violated. This may include sexual harassment, discrimination, wrongful termination, and related misconducts.

For the insurance procured by ANU:

  • The deductible is $100,000 in 2020/21. In (two) prior years, it was $50,000.
  • The upper limit is $2,000,000, in aggregate.

ANU has had a long-term relationship with AIG - Australia for covering this risk. We have AIG data from 2017 through 12 December 2020. There are 11 positive claims for this time frame.

EPL Claims

ANU analysts also have access to additional information such as descriptions of losses. To protect the privacy of claimants, we do not provide details of this information. However, to give you a feel for the kind of information available, the following gives a summary of the claim causes:

  • Alleged unfair dismissal
  • Stop bullying application
  • Dispute regarding employment status
  • General protections dispute
  • Failure to give notice prior to termination
  • Stand down dispute.

3.3.5 Marine Claims

ANU subscribes to two types of marine policies:

  • Marine cargo insurance provides cover for loss or damage to goods in transit by air, sea, road or rail including ocean marine shipments into Australia. This cover applies on a worldwide basis.
  • Marine hull insurance provides cover for loss or damage to marine vessels and equipment and third party liability arising out of the use of the vessels.

Both coverages are with Richard Oliver (QBE). Section 6 provides additional background on these coverages.

Because of the limited experience, no formal analysis of this line has been conducted. In the following, you will find a subset of the experience simply to provide a feel for these type of data.

Marine Hull Claims

3.3.6 Expatriate Insurance

Expatriate insurance policies are designed to cover financial and other losses incurred by expatriates while living and working in a country other than one’s own. To illustrate, ANU currently (2022) has an employee in Papa New Guinea that is covered under its expatriate policy. The ANU policy covers:

  1. medical and additional expenses (upper limit is 1,000,000, annual excess is 250),
  2. medical and emergency evacuation (upper limit is 1,000,000), and
  3. personal liability (upper limit is 1,000,000).

The insurer for this cover is Chubb. The ANU policy began on 16 June 2017 and had renewed each year on 31 December. In the most current year, the policy began on 31 December 2021 and goes to 1 November 2022.

Because of the limited experience, no formal analysis of this line has been conducted. In the following, you will find a subset of the experience simply to provide a feel for these type of data.

ExPat Claims

3.3.7 Cyber / General and Products Liability / Property - Unimutual

The Unimutual data on ANU’s losses for these risks were accessed on 18 August 2020. This database covers Unimutual claims for the period 1 November 2009 to 31 October 2020.

Table 3.12 summarizes the experience including the number of claims and amount incurred by Unimutual, by type of risk. For the cyber claim, the deductible (responsibility of ANU) is $250,000. For the General and Products Liability claims, the deductible is 10,000. The deductible for property risks varied over time and is available in the detailed claims listing below.

Table 3.12: Summary of ANU Unimutual Claims, 2009-2020
Class Number Total Loss Incurred by Unimutual
Cyber 1 1,650,364
General and Products Liability 13 1,083,971
Property 7 312,773,398
Total 21 315,507,733
Unimutual Individual Claims Outcomes


Using the risk classification scheme from Table 3.1, there were no reported Unimutual claims for Professional Indemnity, Medical Malpractice, Clinical Trial, and Crime risk types.

Data access is available in Section 5.5.

4 Workers’ Compensation

Workers’ compensation provides cash and medical benefits to workers who are injured or become ill in the course of their employment and provides cash benefits to the survivors of workers killed on the job.

This risk has several features that warrant considering this coverage separate from other financial risks described earlier in this report. These features include:


4.1 ANU’s Coverage of Workers’ Compensation

In workers’ compensation systems, covered workers are entitled to medical care for their covered injuries or illnesses, and disability benefits to partially replace lost wages. In addition, the survivors of a worker who dies as a result of a covered injury or illness are provided benefits. In general, any injury, illness, or death that arises out of a person’s employment is covered. To get a sense of the size of this obligation, the 2020 ANU annual report (page 97) lists the current liability provision for workers’ compensation to be $2.588 million and the non-current liability provision to be $21.638 million.

For further information, see:

4.2 Sample Workers’ Compensation Data

The data provided to us are maintained by ANU. See the following count of claims.

ANUWCPaid <- read.csv("..\\ANUData\\WCClaims.csv", header = T)
ANUWCPaid$Gross.Paid <- as.numeric(ANUWCPaid$Gross.Paid)
ANUWCPaid$InjuryDate <- as.Date(ANUWCPaid$Injury.Date.B3., "%m/%d/%Y")
ANUWCPaid <- ANUWCPaid[, -1]
rownames(ANUWCPaid) <- NULL
TableWC <- table(format(ANUWCPaid$InjuryDate, format = "%Y"))

knitr::kable(TableWC) %>%
    kableExtra::kable_classic_2(full_width = F, position = "center")
Var1 Freq
2017 20
2018 23
2019 9


ANU analysts have access to additional workers’ compensation information. To protect the privacy of claimants, many of whom are students, we do not provide details of this information. However, to give you a feel for the kind of information available, the following gives a summary.

Nature of Injury/Disease


Distribution of Gross Paid

In the individual claims data, you will find 52 paid claims. The earliest is 2017-01-18 and the latest is 2019-09-09.

To get a sense of the distribution, you will find that the smallest paid claim is 152, the largest is 267992, and the average is 29689.44. The following provides figures to see the distribution of claims. As is common with claims data, the left-hand panel shows the right-skewed nature of the distribution. To help interpret the distribution, the right-hand panel shows the same data but on the logarithmic scale.

**Distribution of Workers Compensation Losses**

Figure 4.1: Distribution of Workers Compensation Losses


As we have done with other lines of business, one can fit a distribution to the losses. In the following, we fit via maximum likelihood the gamma, Pareto, and lognormal distributions to incurred losses. The lognormal distribution appears to be the best fit.

library(VGAM)
x <- seq(0, 15, by = 0.01)

# Inference assuming a gamma distribution
fit.gamma <- vglm(Gross.Paid ~ 1, family = gamma2, data = ANUWCPaid)
theta <- exp(coef(fit.gamma)[1])/exp(coef(fit.gamma)[2])  # theta = mu / alpha
alpha <- exp(coef(fit.gamma)[2])
fgamma_ex <- dgamma(exp(x), shape = alpha, scale = theta) * exp(x)

# Pareto
fit.pareto <- vglm(Gross.Paid ~ 1, paretoII, loc = 0, data = ANUWCPaid)
fpareto_ex <- dparetoII(exp(x), loc = 0, shape = exp(coef(fit.pareto)[2]), scale = exp(coef(fit.pareto)[1])) *
    exp(x)

# Lognormal
fit.LN <- vglm(Gross.Paid ~ 1, family = lognormal, data = ANUWCPaid)
flnorm_ex <- dlnorm(exp(x), mean = coef(fit.LN)[1], sd = exp(coef(fit.LN)[2])) *
    exp(x)

plot(density(log(ANUWCPaid$Gross.Paid)), main = "", xlab = "Log Claims", ylim = c(0,
    0.3))

lines(x, fgamma_ex, col = "blue", lty = 2)
lines(x, fpareto_ex, col = "purple", lty = 3)
lines(x, flnorm_ex, col = "green", lty = 3)
legend("topleft", c("log(Claim)", "Gamma", "Pareto", "Lognormal"), cex = 0.8, lty = c(1,
    2, 3, 4, 1, 1), col = c("black", "blue", "purple", "green"))  #4 in lty is 'longdash'
**Distribution of Workers Compensation Losses with Superimposed Fitted Distributions**

Figure 4.2: Distribution of Workers Compensation Losses with Superimposed Fitted Distributions


Nonetheless, the Pareto distribution is close to the lognormal. For this distribution, the fitted parameters are \(\theta\) (scale) = 30646.22 and \(\alpha\) (shape) = 1.9317. Further, with the fitted distribution, the estimated mean is 32894, the 95th percentile is 113865, and the 99th percentile is 301821.

Analysis by Year

One could also do an analysis of the gross paid distribution by year.

# Analysis By Year
ANUWCPaid.2017 <- subset(ANUWCPaid, format(ANUWCPaid$InjuryDate, format = "%Y") ==
    2017)
ANUWCPaid.2018 <- subset(ANUWCPaid, format(ANUWCPaid$InjuryDate, format = "%Y") ==
    2018)
ANUWCPaid.2019 <- subset(ANUWCPaid, format(ANUWCPaid$InjuryDate, format = "%Y") ==
    2019)

par(mfrow = c(2, 3))
hist(ANUWCPaid.2017$Gross.Paid, main = "", xlim = c(0, 1500000), xlab = "2017 Gross Paid")
hist(ANUWCPaid.2018$Gross.Paid, main = "", xlim = c(0, 1500000), xlab = "2018 Gross Paid")
hist(ANUWCPaid.2019$Gross.Paid, main = "", xlim = c(0, 1500000), xlab = "2019 Gross Paid")
hist(log(ANUWCPaid.2017$Gross.Paid), main = "", xlim = c(5, 16), xlab = "2017 Log Gross Paid")
hist(log(ANUWCPaid.2018$Gross.Paid), main = "", xlim = c(5, 16), xlab = "2018 Log Gross Paid")
hist(log(ANUWCPaid.2019$Gross.Paid), main = "", xlim = c(5, 16), xlab = "2019 Log Gross Paid")
**Distribution of Workers Compensation Losses by Year**

Figure 4.3: Distribution of Workers Compensation Losses by Year

Follow-Up

In the data file, you will find that the gross paid is decomposed into a medical component, payments for incapacity, rehabilitation, and an “other” category (the residual from the gross less the first three components). Some readers will find it interesting to analyze each component on its own.

Further, ANU analysts have access to additional workers’ compensation information. To protect the privacy of claimants, many of whom are students, we do not provide details of this information. However, to give you a feel for the kind of information available, the following gives a summary.

Additional Information Available to ANU Analysts

Data access is available in Section 5.6.

5 Data Dictionary

5.1 SIR Pool Data

Section 2 provides a general introduction to the Self-Insurance Reserve (SIR) Pool Data. Specifically, there are 52 observations in this dataset. The variable names are described in Table 5.1 and the first and last five observations are in Table 5.2. The data are available using this button: .

Table 5.1: Variables in the SIR Pool Dataset
Variable Description
Year Year that the claim occurred
Category Type of claim
Amount Amount of the claim
Description Brief description of the claim
Table 5.2: SIR Pool Data First Five Rows
Year Category Amount Description
2019 Group Personal Accident 257 Claimant tripped on uneven pavement on campus
2019 Group Personal Accident 1435 Bicycle incident - reimbursement for eyewear
2019 Motor Vehicle 700 Car damaged at ANU carpark due to winds
2019 Motor Vehicle 695 Car damaged at ANU carpark due to winds
2019 Motor Vehicle 2368 ANU-owned bollard fall on claimant’s vehicle
Table 5.2: SIR Pool Data Last Five Rows
Year Category Amount Description
2012 Property: Buildings / Contents 21057 Burst Water Pipes
2012 Property: Buildings / Contents 185000 Laboratory Fire
2012 Property: Buildings / Contents 85622 Cracked Glass Furnace
2012 Public Liability 58333 Liability Claimant
2012 Public Liability 25000 Liability Claimant


Source: Frees, Edward and Butt, Adam (2022). “ANU Self Insurance Reserve Pool Losses 2022”. Australian National University Data Commons. DOI https://doi.org/10.25911/74sx-x144.

5.2 Corporate Travel Data

Section 3.3.1 provides a general introduction to the Corporate Travel Data. There are 2107 observations in this dataset. The variable names are described in Table 5.3 and the first and last five observations are in Table 5.4.

Data are available using this button: .

Table 5.3: Variables in the Corporate Travel Dataset
Variable Description
UW Year Underwriting Year
Loss Date Date that the loss occurred
Reported Date Date that the loss was reported
Last Trans Date Last date in which there was a transaction regarding the loss
Paid Loss Cumulative amount paid on the loss
Outstanding Reserve Estimate of the loss amount yet to be paid
Incurred Loss Sum of the amount paid and the estimate of future payments
Status An indicator as to whether the claim has been deemed settled (closed) or not settled (open)
Table 5.4: Corporate Travel Data First Five Rows
UW.Year Loss.Date Reported.Date Last.Trans.Date Paid.Loss Outstanding.Reserve Incurred.Loss Status
2021 19/12/2021 20/12/2021 24/12/2021 10000.00 0 10000.00 Closed
2021 9/4/2022 29/04/2022 30/05/2022 423.08 0 423.08 Closed
2021 2/5/2022 4/5/2022 0.00 500 500.00 Open
2021 5/5/2022 17/05/2022 0.00 562 562.00 Open
2021 30/04/2022 27/05/2022 10/6/2022 1500.00 0 1500.00 Closed
Table 5.4: Corporate Travel Data Last Five Rows
UW.Year Loss.Date Reported.Date Last.Trans.Date Paid.Loss Outstanding.Reserve Incurred.Loss Status
2006 1/11/2006 19/06/2007 0.00 0 0.00 Closed
2006 24/06/2007 26/06/2007 8/1/2008 6278.10 0 6278.10 Closed
2006 4/7/2007 6/7/2007 11/9/2007 114.50 0 114.50 Closed
2006 20/05/2007 26/06/2007 14/07/2007 135.65 0 135.65 Closed
2006 15/02/2007 27/06/2007 14/07/2007 1207.75 0 1207.75 Closed


Source: Frees, Edward and Butt, Adam (2022). “ANU Corporate Travel Insurance Claims 2022”. Australian National University Data Commons. DOI https://doi.org/10.25911/vrdw-9f32.

5.3 Group Personal Accident Data

Section 3.3.2 provides a general introduction to the Group Personal Accident Data. There are 148 observations in this dataset. The variable names are described in Table 5.5 and the first and last five observations are in Table 5.6.

Data are available using this button: .

Table 5.5: Variables in the Group Personal Accident Dataset
Variable Description
UW Year Underwriting Year
Loss Date Date that the loss occurred
Last Trans Date Last date in which there was a transaction regarding the loss.
Paid Loss Cumulative amount paid on the loss
Outstanding Reserve Estimate of the loss amount yet to be paid
Incurred Loss Sum of the amount paid and the estimate of future payments
Status An indicator as to whether the claim has been deemed settled (closed) or not settled (open)
Table 5.6: Group Personal Accident Data First Five Rows
UW.Year Loss.Date Last.Trans.Date Paid.Loss Outstanding.Reserve Incurred.Loss Status
2021 6/12/2021 3/6/2022 805.0 0.0 805 Closed
2021 15/11/2021 0.0 0.0 0 Closed
2021 15/11/2021 0.0 0.0 0 Closed
2021 22/03/2022 4/5/2022 396.0 0.0 396 Closed
2021 11/4/2022 2/8/2022 740.1 359.9 1100 Open
Table 5.6: Group Personal Accident Data Last Five Rows
UW.Year Loss.Date Last.Trans.Date Paid.Loss Outstanding.Reserve Incurred.Loss Status
2010 6/3/2011 26/07/2011 776.00 0 776.00 Closed
2010 22/07/2011 23/01/2012 4624.54 0 4624.54 Closed
2010 5/6/2011 30/01/2012 1503.65 0 1503.65 Closed
2007 11/1/2008 23/02/2008 0.00 0 0.00 Closed
2007 29/08/2008 0.00 0 0.00 Closed


Source: Frees, Edward and Butt, Adam (2022). “ANU Group Personal Accident Claims 2022”. Australian National University Data Commons. https://doi.org/10.25911/jcfx-zj56.

5.4 Motor Vehicle Data

Section 3.3.3 provides a general introduction to the Motor Vehicle Data. There are 318 observations in this dataset. The variable names are described in Table 5.7 and the first and last five observations are in Table 5.8.

Data are available using this button: .

Table 5.7: Variables in the Motor Vehicle Dataset
Variable Description
Policy Term Start Date Start date of the contract year in which the loss occurred
Loss Date Date that the loss occurred
Reported Date Date that the loss was reported
Motor Fault Party responsible for the loss
Driver Age Age of the driver
Vehicle Description Type of vehicle
Loss Postcode Postal code where the loss occurred
Excess The deductible applied to the loss
Motor Net Paid Amount paid to the insured (ANU)
Outstanding Estimate Estimate of the loss amount yet to be paid
Motor Net Incurred Sum of the amount paid and the estimate of future payments
Third Party Identified Indicates whether a responsible third party could be identified
Third Party Insured Indicates whether a responsible third party was insured
Table 5.8: Motor Vehicle Data First Five Rows
Policy.Term.Start.Date Loss.Date Reported.Date Motor.Fault Driver.Age Vehicle.Description Loss.Postcode Excess Motor.Net.Paid Outstanding.Estimate Motor.Net.Incurred Third.Party.Identified Third.Party.Insured
1/11/2011 6/6/2012 4/10/2012 THIRD PARTY RESPONSIBLE NA FORD TRANSIT VAN 2600 1000 384.88 0 384.88 IDENTIFIED
1/11/2011 16/08/2012 14/11/2013 INSURED RESPONSIBLE 39 TOYOTA HIACE 2612 1000 901.21 0 901.21
1/11/2011 4/9/2012 17/01/2013 INSURED RESPONSIBLE 52 HYUNDAI IX35 2600 1000 1225.71 0 1225.71
1/11/2011 21/09/2012 28/09/2012 THIRD PARTY RESPONSIBLE 59 HOLDEN COMMODORE 2518 NA 1671.76 0 1671.76 IDENTIFIED NOT INSURED
1/11/2011 22/09/2012 12/10/2012 INSURED RESPONSIBLE NA SUBARU FORESTER 2612 1000 3418.86 0 3418.86 INSURED
Table 5.8: Motor Vehicle Data Last Five Rows
Policy.Term.Start.Date Loss.Date Reported.Date Motor.Fault Driver.Age Vehicle.Description Loss.Postcode Excess Motor.Net.Paid Outstanding.Estimate Motor.Net.Incurred Third.Party.Identified Third.Party.Insured
1/11/2021 4/4/2022 5/4/2022 INSURED RESPONSIBLE 66 VOLKSWAGEN TIGUAN 2604 0 2373.49 1056.00 3429.49
11/1/2021 11/4/2022 9/5/2022 INSURED RESPONSIBLE 27 TOYOTA HILUX 2540 0 210.00 25000.00 25210.00
1/11/2021 11/4/2022 9/5/2022 INSURED RESPONSIBLE 27 TOYOTA HILUX 2540 0 0.00 31927.27 31927.27
11/1/2021 15/04/2022 11/7/2022 INSURED RESPONSIBLE 21 TOYOTA HILVX 2601 0 0.00 2750.00 2750.00
1/11/2021 18/07/2022 18/07/2022 NO-ONE RESPONSIBLE NA TOYOTA HILUX 2601 0 0.00 299.00 299.00


Source: Frees, Edward and Butt, Adam (2022). “ANU Motor Vehicle Claims 2022”. Australian National University Data Commons. DOI https://doi.org/10.25911/g7e4-9e46.

5.5 Unimutual Claims Data

Section 3.3.7 provides a general introduction to the Unimutual Claims Data. There are 21 observations in this dataset. The variable names are described in Table 5.9 and the first and last five observations are in Table 5.10.

Data are available using this button: .

Table 5.9: Variables in the Unimutual Claims Dataset
Variable Description
Class Type of loss
Year Underwriting year in which the loss occurred
Incident Type Brief description of the cause of the loss
Retention The amount that the member (ANU) is responsible for
NetPaid The amount paid by the (Unimutual) fund
Outstand Estimate of the loss amount yet to be paid
NetLessMember The amount paid by the (Unimutual) fund minus the amount paid by the member (ANU)
Table 5.10: Unimutual Claims Data First Five Rows
Class Year Incident.Type Retention NetPaid Outstand NetLessMember
Property 2009 - 2010 Fire/Explosion 750000 432282.02 0 232282.02
Property 2011 - 2012 Fire/Explosion 200000 2737242.73 0 2537242.73
General and Products Liability 2011 - 2012 Slip/Trip/Fall 10000 7522.73 0 0.00
Property 2012 - 2013 Fire/Explosion 200000 11182841.63 0 10982841.63
Property 2012 - 2013 Water Damage 200000 10481.64 0 9981.64
Table 5.10: Unimutual Claims Data Last Five Rows
Class Year Incident.Type Retention NetPaid Outstand NetLessMember
Cyber 2018 - 2019 Breach of Privacy 250000 33093.71 1867270.29 1650364
General and Products Liability 2018 - 2019 Bodily Injury (excl Slip, Trip & Fall) 10000 2471.00 0.00 0
Property 2019 - 2020 Storm 500000 56608208.68 193391791.30 249500000
General and Products Liability 2019 - 2020 Libel/Slander 10000 8617.50 1382.50 0
General and Products Liability 2019 - 2020 Libel/Slander 10000 5375.50 4624.50 0


Source: Frees, Edward and Butt, Adam (2022). “ANU Unimutual Claims 2022”. Australian National University Data Commons. DOI https://doi.org/10.25911/ymnw-6a81.

5.6 Workers Compensation Data

Section 4 provides a general introduction to the Workers Compensation Data. There are 52 observations in this dataset. The variable names are described in Table 5.11 and the first and last five observations are in Table 5.12.

Data are available using this button: .

Table 5.11: Variables in the Workers Compensation Dataset
Variable Description
Injury Date(B3) Date that the injury occurred
Claim status code Claim status code
Nature of injury disease Nature of injury disease
Gross Est Gross estimate
Gross Paid Gross paid
Medical PTD Medical paid to date
Incapacity PTD Incapacity paid to date
Rehab PTD Rehabilitation paid to date
Table 5.12: Workers Compensation Data First Five Rows
Injury.Date.B3. Claim.status.code Nature.of.injury.disease Gross.Est Gross.Paid Medical.PTD Incapacity.PTD Rehab.PTD
2/8/2018 C 1 186462 127371 16587 93908 11755
4/12/2018 C 1 68150 76647 5520 66522 0
5/31/2017 F 1 0 76270 12017 45507 18161
7/25/2017 F 1 0 34070 8417 13862 8957
3/1/2018 F 1 0 25430 2649 19955 1230
Table 5.12: Workers Compensation Data Last Five Rows
Injury.Date.B3. Claim.status.code Nature.of.injury.disease Gross.Est Gross.Paid Medical.PTD Incapacity.PTD Rehab.PTD
11/23/2017 C 23 90127 66916 16812 32033 17807
11/24/2018 F 23 0 1741 1741 0 0
9/9/2019 C 24 17706 12294 9245 0 2978
6/12/2018 F 24 0 30280 7473 22497 0
1/30/2018 F 24 0 12638 12638 0 0


Source: Frees, Edward and Butt, Adam (2022). “ANU Workers Compensation Losses 2022”. Australian National University Data Commons. DOI https://doi.org/10.25911/y8a3-t990.

6 Appendix. Description of ANU Risks Covered by Insurance

Descriptions of these insurance classes are broadly available, see, for example, the International Risk Management Institute.

Property

The property protection cover is a large portion of the ANU risk portfolio. It is a complex coverage that cannot easily be represented using a simple deductible, upper limit, and premium. This cover is subdivided into two main sections:

  • Material loss or damage that covers real and personal property in which ANU has an insurable interest. For the 2020/2021 policy year, this is limited by ANU’s physical assets that were approximately 4.5 billions of AUD.
  • Consequential loss that covers other damages suffered by the insured, ANU, resulting from the interruption or interference of business. For the 2020/2021 policy year, this is limited by a contractually agreed upon limit, 1 billion AUD.
  • For both material and consequential loss, there is a combined upper limit of 500 millions of AUD for losses due to natural catastrophes.

Within these two broad sections, there is a host of sub-limits of liability (that are applied in excess of the deductible). These limits are for a single loss, or a series of losses related to a single event. For material loss or damage, they include accidental damage (25 million), burglary and theft (1 million), replacement of locks and keys (2 millions), and so on. For consequential loss, they include loss of revenue or increased cost of work (60 millions), termination of employment expenses (19 millions), and so on.

General and Products Liability

General and products liability insurance protects businesses from most liability exposures other than automobile and professional liability. It can include bodily injury and property damages caused to others. Or, it may be liability for financial damages that result from libel, slander, wrongful eviction or false arrest, or violating one’s right to privacy. Some insurers sell separately product liability insurance that provides protection against financial loss arising out of the legal liability incurred by an insured because of injury or damage resulting from the use of a covered product.

Cyber

ANU’s cyber insurance provides a cover for losses arising out of a breach of personal as well as corporate information (including claims against an outsourcer), data security liability, media content liability, cyber extortion, network interruption and defense costs.

Broadly, this policy covers a variety of both liability and property losses that may result when ANU engages in various electronic activities. For example, it covers liability for a data breach in which students’ personal information is exposed or stolen by a hacker or other criminal. It can cover a variety of expenses associated with data breaches including notification costs, credit monitoring, costs to defend claims, fines and penalties, and loss resulting from identity theft.

Crime

Crime insurance provides ANU protection from employee and executive fraud or dishonesty, third-party crime, as well as electronic and computer crime. In general, it provides coverage for loss of money, securities, or other assets resulting from acts such as employee theft, certain types of fraud by third parties (forgery, for example), theft of property from the premises, and social engineering (impersonation fraud). That is, crime insurance provides cover for the loss of property (money or goods) belonging to ANU directly resulting from the dishonest acts committed by an employee for his or her own personal gain whether acting alone or in collusion with others.

Employment Practices Liability

Employment practices liability is a type of liability for wrongful acts arising from the employment process. The insurance covers claims by workers that their legal rights as employees of the company have been violated. This may include sexual harassment, discrimination, wrongful termination, and related misconducts.

At ANU, the claim causes include:

  • Alleged unfair dismissal
  • Stop bullying application
  • Dispute regarding employment status
  • General protections dispute
  • Failure to give notice prior to termination
  • Stand down dispute

Expatriate Insurance

Expatriate insurance policies are designed to cover financial and other losses incurred by expatriates while living and working in a country other than one’s own. The most common insurance policies purchased by expatriates include:

  • Personal property
  • Automobile insurance
  • Personal liability insurance
  • Emergency evacuation
  • Medical and dental coverage
  • Short-term travel insurance

Group Personal Accident

Group personal accident insurance offers financial protection in case of injury or death resulting from an incident that occurs on the job. Like workers’ compensation, group personal accident offers insurance coverage and liability insurance protection against accidental death or injury. Unlike workers’ compensation, group personal accident covers students and ANU’s voluntary workers. Unlike workers compensation Insurance, the cover applies 24 hours a day, 7 days a week. Further, the cover applies when traveling on official university business; so, it can be arranged to apply worldwide.

The policy provides a lump sum benefit for an injury that results directly in death, permanent total disablement, paraplegia, and quadriplegia, loss of sight, loss of limb or limbs and loss of hearing. In addition and subject to any sub limits the policy provides cover for the following expenses incurred as a result of an injury:

  • non-Medicare Medical expenses
  • Emergency Home Help
  • Student Tutorial Costs
  • Out of Pocket expenses
  • Emergency Transport Expenses
  • Aids/HIV Extension

Corporate Travel

Universities purchase corporate travel policies to cover employees and students traveling on official university business for a wide variety of accidents and incidents while away from the campus or primary workplace. This broad coverage includes medical care and evacuation, loss of personal property, extraction for political and weather related reasons, and more. According to ANU’s insurer, Chubb Travel, this insurance can cover:

  • Medical, evacuation and additional expenses
  • Cancellation and disruption
  • Baggage and travel documents
  • Political and natural disaster evacuation
  • Accidental death and disability
  • Rental and personal vehicle excess
  • Alternative employee/Resumption of assignment
  • Kidnap and ransom/Extortion
  • Hijack and detention
  • Personal liability
  • Extra territorial workers compensation
  • Search and rescue expenses

Professional Indemnity, Medical Malpractice, and Clinical Trials

Professional indemnity, medical malpractice, and clinical trials are types of professional liability coverages that are designed to protect businesses under the complaint that they were harmed by a professional’s negligence or intentionally harmful treatment decisions.

Deductibles and policy upper limits vary by line. For ANU, these include:

  • Claims Preparation Costs 25,000
  • Compensation for Court Attendance 50,000
  • Emergency Defence Costs 100,000
  • Inquiry Costs 1,000,000
  • Loss of Documents 2,000,000
  • Public Relations Expenses 100,000
  • Sexual Misconduct Defence Costs 100,000
  • Statutory Liability Costs and Penalties 250,000
Professional Indemnity

Professional indemnity insurance is designed to protect business owners, freelancers and the self-employed if clients claim a service is inadequate. Any organization that provides a professional service or gives advice could be sued if the recipient is unhappy with their work.

Professional indemnity typically provides coverage on a claims made basis which means that claims first notified and reported to the insurer during the period of insurance are recoverable. The actual loss date does not matter, subject to the provisions of any clause relating to a retroactive date. In contrast, an occurrence policy covers claims that occurred while the policy was in effect.

To be specific, ANU’s Professional indemnity excess/deductible is 100,000. The Professional Indemnity upper policy limit is $20,000,000 plus two reinstatements.

  • “Reinstatement” — under many forms of reinsurance and insurance, the payment of a claim reduces an aggregate limit by the amount of the claim. Provision is sometimes made for reinstating the policy limit to its original amount when the original limit has been exhausted.
Medical Malpractice

Medical malpractice, also known as medical professional liability, is a type of insurance that provides compensation to injured patients and families because of health care provider negligence, see for example, Frees and Gao (2020). It covers the acts, errors, and omissions of physicians and surgeons, encompassing physicians professional liability insurance, hospital professional liability (HPL) insurance, and allied healthcare (e.g., nurses) professional liability insurance.

Although the majority of policies are written with a claims-made coverage trigger, such coverage is sometimes available on an occurrence basis.

Clinical Trials

Clinical trials insurance provides protection for the organizers of clinical trials for drug and medical device testing. It covers their legal liability to pay compensation in the event of an injury to a trial participant. Such liability can arise in all phases of clinical trials for drugs and medical devices including negligent harm to trial participants and non-negligent harm (also known as no-fault compensation). The underwriting for this insurance is based on an agreed protocol and informed patient consent form that describes the objectives, design, methodology, statistical considerations, and organization of a clinical trial.

Statutory Liability

As summarized in (“Statutory liability” 2022), businesses are responsible for complying with a myriad of local, state, and federal laws and regulations. Accidental breaches of the law can put a company at risk for payments in lawsuits, compensatory damages, and settlements to resolve claims. In Australia, businesses commonly purchase statutory liability insurance to protect themselves from the fines, penalties, and legal fees that can result from an accidental breach of law. These may include occupational health and safety laws, environmental laws, and employment laws. This insurance coverage can include expenses for defense costs, inquiry costs, fines, and penalties, where insurable by law.

Not surprisingly, breaches of Work, Health and Safety legislation account for a high proportion of claims being made under Statutory Liability. Measures are being taken to curb workplace accidents in those industries with a higher element of danger including Mining, Agriculture and Construction. However, due to the physically intense nature of the work, these employees face a greater risk of injury than those in less active roles.

Motor Vehicle

This policy covers ANU’s vehicles including cars, vans, utilities, and motorcycles. There are two parts to this coverage, one for comprehensive damage to the insured vehicles and a second for legal liability.

The first part is for loss or damage to insured vehicles, including:

  • Market Value for Cars, vans, utilities, motorcycles, or 4WD’s of less than two tonne carrying capacity.
  • All other vehicle types: Market Value or the amount shown in the agreed schedule of vehicles, whichever is less.

The second part is for legal liability arising from claims from a third party. This cover is limited by:

  • 50 Million for all claims arising from the one accident or series of accidents resulting from the one original cause.
  • 1 Million for all claims arising from the one accident or series of accidents resulting from the one original cause where the vehicle is used for transportation of dangerous goods.

Marine Cargo and Hull Coverages

ANU subscribes to two types of marine policies:

  • Marine cargo insurance provides cover for loss or damage to goods in transit by air, sea, road or rail including ocean marine shipments into Australia. This cover applies on a worldwide basis.
    • The insurable interest consists of all goods and/or merchandise of every description for which the Insured is responsible, consisting principally of, but not limited to scientific and research equipment, including returned and/or reshipped and/or re-consigned Interest. Other interests held covered on cover conditions at rates to be agreed.
    • There is a 5,000 deductible on each and every claim.
    • There is a 5 million upper limit on claims.
  • Marine hull insurance provides cover for loss or damage to marine vessels and equipment and third party liability arising out of the use of the vessels.
    • There is a 150 deductible on each and every claim, excluding all third party liability claims.
    • There is a 5 million upper limit on third party liability claims.
    • This policy covers the insured’s unpowered craft whilst being used during training, racing, sporting competitions and all other associated activities.
    • In 2020, the sum insured is 1,042,209 for hull, machinery, fittings and equipment, mast, sails and rigging of insured vessels as declared as:
      • Boat Club Asset schedule 574,478
      • Scuba Club 205,305
      • Mountaineering Club 77,191
      • Boat Shed 72,410
      • Sailing Club 112,825.

Citation

For attribution, please cite this work as

Frees, Edward W. and Butt, Adam (2022). “ANU insurable risks.” Australian National University Open Research Library. https://doi.org/10.25911/0SE7-N746.


Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

References

Actuarial Community (2020), Loss data analytics. Accessed https://openacttexts.github.io/Loss-Data-Analytics/index.html. Retrieved on October 15, 2020.

ANU (2018), “ANU self-insurance reserve guidelines.” Accessed https://services.anu.edu.au/files/guidance/SIR%20Policy%202018_2019.pdf. Retrieved on 15 May, 2022.

——— (2022a), “ANU risk management policy statement.” Accessed https://policies.anu.edu.au/ppl/document/ANUP_000462. Retrieved on 15 May, 2022.

——— (2022b), “ANU insurance policy statement.” Accessed https://policies.anu.edu.au/ppl/document/ANUP_000422. Retrieved on 15 May, 2022.

Frees, Edward W and Lisa Gao (2020), “Predictive analytics and medical malpractice,” North American Actuarial Journal, 24 (2), 211–27.

Insights, Deloitte (2019), “Global risk management survey, 11th edition,” London, England, Deloitte Insights. Accessed https://www2.deloitte.com/content/dam/Deloitte/co/Documents/risk/DI_global-risk-management-survey.pdf. Retrieved on May 27, 2020.

Solutions, Aon Risk (2019), “Global risk management survey,” London, England, Aon Risk Solutions. Accessed https://www.aon.com/2019-top-global-risks-management-economics-geopolitics-brand-damage-insights/index.html. Retrieved on May 27, 2020.

“Statutory liability” (2022). Accessed https://www.investopedia.com/terms/s/statutory-liability.asp. Retrieved on 15 May, 2022.